Statistical Modelling of Oil Expression From Neem Seed Using A Screw Press

Arid Zone Journal of Engineering, Technology and Environment, September, 2018; Vol.
14(3):450-468
Copyright © Faculty of Engineering, University of Maiduguri, Maiduguri, Nigeria.
Print ISSN: 1596-2490, Electronic ISSN: 2545-5818, www.azojete.com.ng
STATISTICAL MODELLING OF OIL EXPRESSION FROM NEEM SEED USING A

SCREW PRESS
1
*Orhevba, B. A., 2Chukwu, O., 3Zinash, D. and 4Sunmonu, M. O.
1, 2,3
Department of Agricultural and Bioresources Engineering, Federal University of Technology, P.M.B
65, Minna, Niger State, Nigeria.
4
Department of Food Engineering, University of Ilorin, Ilorin
*Corresponding author’s Email: borhevba@yahoo.com
Abstract
A study was carried out to determine the effect of some process parameters on the yield of neem seed oil
from a screw press using 44 factorial design experiments. This consisted of four independent, randomly
assigned parameters at four levels each namely, moisture content (wb) of seeds (6.3%; 8.1%, 13.2% and
16.6%); temperature of heating at ϴ1 (55°C), ϴ2 (70°C), ϴ3(85°C) and ϴ4 (100°C); duration of heating at t1
(5 min.), t2(10 min.), t3 (15 min.) and t4 (20 min.) and machine pressure at p1 (602.91 kN/m2), p2 (723.07
kN/m2), p3(925.84 kN/m2) and p4(1334.88 kN/m2). Three replicates of the experiments were carried out and
from the values obtained, oil yield was calculated. The effects of the processing conditions (moisture
content, heating duration, heating temperature and machine pressure) on the oil yield were investigated
using analysis of variance (ANOVA) at p≤0.05 and the levels of the significant means were further
evaluated using Duncan’s New Multiple Range Test (DNMRT). Mathematical model was developed to
express the oil yield with respect to the process parametersand this was fitted to experimental data using
MATLAB 8.0 software package. Adequacy of the model was authenticated using coefficient of
determinationand adjusted coefficient of determination, R2.The results obtained showed that temperature,
heating time, moisture content, pressure and their interactions all proved to determine the outcome of oil
yield of neem seed. The highest oil yield of 33.55% was obtained at temperature of 81.98°C, pressure of
1067.23kN/m2, moisture content of 9.03% and heating duration of about 20 min.Theresults
obtained(coefficient of determination and adjusted coefficient of determination R2) indicated that the model
generated was statistically adequate.
Keywords: Neem seed, Oil, Factorial design, Moisture content, Temperature
1. Introduction
The neem (Azidirachtaindica) tree is popularly known as dongoyaro in the Northern part of
Nigeria where it grows in abundance. All parts of the tree have been reported to be very useful
(Adewoye and Ogunleye, 2012). They also reported that the most famous product of the tree is the
oil obtained from the seed kernel. Neem seed kernel is a component of the neem fruit which has
high concentration of oil (Ikasari and Indraswati, 2008), Ahmed and Grainge (1985) reported that
neem seed contains 35 – 45% oil. The quality of the oil differs according to the method of
processing.
Oil extraction is the process of expelling oil from oil bearing agricultural seeds and there are
different methods employed in the extraction process. These include the traditional method,
supercritical fluid extraction method, mechanical method and the solvent extraction process
(contact equilibrium process) or a combination of mechanical and solvent extraction processes
(Oyinlola and Adekoya, 2004). Oil extraction by mechanical pressing is simpler, safer and
contains fewer steps, compared to oil extraction by solvent (Oyinlola and Adekoya, 2004).
Mechanical expression of oil involves the application of pressure (using hydraulic or screw
presses) to force oil out of the oil bearing material.
450
Arid Zone Journal of Engineering, Technology and Environment, September, 2018; Vol. 14(3):450-468.
ISSN 1596-2490; e-ISSN 2545-5818; www.azojete.com.ng
Heat treatment of oil seeds has been observed to rupture the oil bearing cells of the seed, coagulate
the protein in the meal, adjust the moisture level of the meal to optimum level for oil expression,
lower the viscosity and increase the fluidity of the oil to be expelled, thereby facilitating oil
expression from the material (Adeeko and Ajibola, 1990). Effect of different processing factors on
yield and quality of oil from some oil bearing seeds have also been investigated by various authors.
These include groundnut (Adeeko and Ajibola, 1990); olive (Torres and Maestri, 2006); sesame
seed (Akinoso et al., 2006a), palm kernel (Akinoso, 2006); soyabean (Tunde-Akintunde et al.,
2001); conophor nut (Fasina and Ajibola, 1989), castor seed (Shridhar et al., 2010), and neem seed
(Adewoye and Ogunleye, 2012). Adewoye and Ogunleye (2012) used the response surface
methodology to optimize oil extraction from neem seed.
Neem oil is used as a base for variety of organic cosmetics including soaps, shampoos, hand and
body lotions and creams (Rajev, 2009). It is also used as an organic bio-pesticide repellant against
insects such as meal worms and aphids (Rajev, 2009). Abdullahi (2004) reported thatneem oil is
used for treating many skin diseases such as eczema, psoriasis (skin disease) and skin allergies.
Modelling reduces number of experiments; thereby reducing time and expenses and providing
process optimization, predictive capability, improved process automation and control possibilities
(Sablani et al. 2006) (The yield of the oil obtained by mechanical expression are affected by
various operating conditions such as heating temperature, heating time, moisture content, applied
pressure, particle size and pressing time (Khan and Hanna, 1983). According to Bamgboye and
Adejumo (2011), for maximum oil yield and least residual oil in cake, it is very important to
control these conditions during the extraction process. Thus, for efficient mechanical expression, a
careful establishment of optimum processing conditions is necessary) The intention of this study
therefore is to determine the effect of moisture content, temperature of heating, duration of heating
and pressure on the yield of neem seed kernel oil expressed using an oil expeller and to develop
model equations to optimise oil expression from neem seed kernels.
2.0 Materials and Methods
2.1 Materials
Mature and healthy neem seeds used for the experiments were obtained from Katsina Zonal Forest
Office, Ministry of Agriculture and Natural Resources, Katsina, Katsina State, Nigeria. A neem
seed roaster which uses steam as a source of heat energy was used in heating the neem seeds prior
to oil expression. NCRI, Badeggi developed oil expeller was used in carrying out oil expression.
The expeller capacity ranged from 15 – 20 kg/h and was powered by a 7.5 kW, 3 phase electric
motor with in-built reduction gear. It was run at 75 rpm.
Fig 1: NCRI, Badeggi Developed Oil Expeller

451
Orhevba et al.: Statistical Modelling of Oil Expression from Neem Seed using a Screw Press. AZOJETE,
14(3):450-468. ISSN 1596-2490; e-ISSN 2545-5818, www.azojete.com.ng
2.2 Methods
The following procedures were used in carrying out the experiments.
2.2.1 Material Preparation
The seeds still within their endocarp were sun-dried for two days to allow for their easy removal.
The dried endocarp was cracked to obtain the seed kernel and after decortication, the hulls of the
seeds and other contaminants were removed by winnowing.
2.2.2 Moisture Content Determination

The moisture content of the seeds was determined using ASABE (2008) standard for oil seeds.
Three samples each weighing 15 g were placed in an oven set at 105°C and dried for 24 hours. The
samples were then cooled, weighed and the moisture content calculated. Loss in weight was
assumed to be moisture loss. Initial moisture content of the seeds was 8.1%. The sample was
divided into four parts; one part (one-fourth) was left as it was; while the remaining parts were
sun-dried at 34°C for 12 hours to further reduce the moisture content to 6.3%. This sample was
further divided into three parts, one part (one-third) was left as it was; while the remaining parts
were further conditioned to desired moisture content levels as described by Akinoso (2006).
Adding distilled water as calculated from equation 1 increased the moisture content of the seeds
according to Akinoso (2006)
(1)
where:
A = Initial mass of the sample
a = Initial moisture content of the sample, % wb
b = Final (desired) moisture content of sample % wb
Q = Mass of water to be added kg
Each sample was sealed in a separate polythene film. The samples were kept at 5°C in a
refrigerator for a week to enable the moisture to distribute uniformly throughout the samples. The
four moisture content levels that were prepared are: 6.3% wb, 8.1% wb, 13.2% wb and 16.6% wb.
Based on literature search, I decided to go below and slightly above the values seen.
2.2.3 Heating of Seeds
A seed roaster of 15kg/hr capacity, which uses steam as a source of heat energy was used in
heating the neem seeds prior to expelling of oil from them. The already conditioned seeds were
poured in the fabricated seed roaster which had been turned on before now. As soon as the
required temperature was reached, the samples were poured inside for the required heating
duration. A thermometer (which measures the temperature of the seeds being heated) was attached
to the seed roaster. Neem kernel samples of 2 kg each were heated in the seed roaster at different
temperature and time combinations. A stopwatch was used to monitor the time. As soon as the
specified heating duration was reached, the seeds were discharged/released from the roaster
through its seed discharge outlet directly into the hopper of the oil expeller.
2.2.4 Determination of Machine Pressure
Preliminary testing was done to determine (calculate) the pressure generated within the system:
the diameter of the barrel of the oil expeller was a constant and this was known; worm was
loosened completely, the worm shaft had four different points with different diameters (measured,
noted and marked), these points on the worm correspond to other points on the shank (outer part of
452
the worm shaft which could be seen even when the worm shaft is inside the barrel) which were
already marked, these points on the shank could be seen as the shaft was adjusted. The more the
worm shaft was adjusted inwards through the adjustment mechanism, the higher the diameter
covered by the worm and the lesser the clearance between the worm and the barrel and thus the
more the pressure generated within the system. The measured diameters at the different points on
the worm shaft are 0.07, 0.072, 0.074 and 0.076m. These values were substituted into the formula
reported by Hannah and Stephens (1984) to obtain the Torque (equation 2).
(2)
Where:
PW = Power, Watts
N= Revolutions per minute
T= Torque, Nm
Torque = Force × Distance (3)
The torque and the diameters calculated were then substituted into equation 2 to get the different
pressures exerted by the machine.
Also,
(4)
Where:
Pr= Pressure, N/m2
F = Force, N
A= Area, m2
The applied pressures calculated were 602.91, 723.07, 925.84 and 1,334.88 kN/m2, respectively
2.2.5 Oil Expression
Neem seed oil was expressed using NCRI, Badeggi multi-seed screw press (Fig.1).The expeller
capacity ranged from 15–20 kg/h and was powered by a 7.5 kW, 3 phase electric motor with in-
built reduction gear. It was run at 75 rpm.
The experimental procedure was carried out after running the screw press for about 3 min before
loading the pre-treated samples as described by Akinoso (2006). Oil expressions were conducted at
moisture contents of 6.3, 8.1, 13.2 and 16.6% wb. They were heated at 55, 70, 85 and 100°C at 5,
10, 15 and 20 min heating duration using a steam roaster. The applied pressure was 602.91,
723.07, 925.84 and 1,334.88kN/m2, respectively. Oil expressed and the cakes from the samples
were collected separately. Cleaning of the expeller barrel was done after each expression. The
experiments were replicated three times. The data obtained for oil yield was regressed using
multiple regression analyses in MATLAB (8.0) computer software.
2.2.6 Determination of Oil Yield
The expressed oil was collected and left to stand for 96 hours as recommended by Olajide (2000)
so that the oil can be clarified and the volume measured. The weights of the cakes were determined
using an electronic weighing balance. The recorded oil yield was compared with the initial oil
content of the seed.
453
Soxhlet oil extraction method as reported by Akinosoet al. (2006a) was applied to determine the
initial oil content of the seeds. The mean value of three samples was expressed as percentage
content as follows:
(5)
The weights of the oils expressed were obtained using a weighing balance. Percentage oil yield
and Expression efficiency were calculated using equations (1) and equation (2) respectively.
(6)
(7)
2.3 Statistical Analysis

A 44 full factorial design was used for the experimentation. Four parameters were measured at four
levels each which gave 256 treatment combinations. These parameters were moisture content of
neem seeds, machine pressure, duration and temperature of heating; and these were independent
variables while oil yield was the dependent variable. The parameters and their levels are shown in
Tables 1. The choice of moisture content of seeds, machine pressure, duration and temperature of
heating were influenced from literature.
The necessary parameters were varied taking into consideration the reported values for oil seeds.
Obtained data from these experiments formed the basis for the model equation. The parameters m1,
m2, m3, m4 are moisture contents at 6.3, 8.1, 13.2 and 16.6 %, respectively; ϴ1, ϴ2, ϴ3, ϴ4 are
temperatures at 55, 70, 85 and 100°C, respectively; p1, p2, p3, p4 are machine pressures at 602.834,
743.628, 978.628 and 1,449.12kN/m2, respectively while t1, t2, t3, t4 are heating durations at 5, 10,
15, and 20 minutes, respectively (Table 1).All the treatments were randomly assigned.
Table 1: Parameters and their Levels
Parameters Levels
1 2 3 4
Moisture Content % wb , m 6.3 8.1 13.2 16.6
Heating Temperature °C, ϴ 55 70 85 100
Heating Duration (min), t 5 10 15 20
Machine Pressure (KN/m2), p 602.91 723.07 925.84 1334.88
The effect of the processing conditions (moisture content, heating temperature, heating duration
and machine pressure) on yield of the oil was investigated using analysis of variance (ANOVA) at
p≤0.05 and levels of the significant means were further evaluated using Duncan’s New Multiple
Range Test (DMRT).
The data was analysed using multiple regression technique of the MATLAB (8.0) software to
generate a mathematical model. Adequacy of the model generated was authenticated by the
coefficient of determination and adjusted coefficient of determination R2.
2.4 Model Development
All the possible relationships between the process conditions manipulated (temperature, pressure,
moisture content and heating duration) and the measured output (oil yield) were explored with a
view to selecting models that appropriately capture the relationship between the input and output
454
parameters. The model development process therefore took cognizance of the various relationships
exhibited between the process conditions and the output.
The data obtained for oil yield was regressed using multiple regression analyses in MATLAB (8.0)
computer software. The essence is to find functional relationship that can adequately relate process
parameters (temperature, pressure, heating duration and moisture content) to the measured output
(oil yield) Boonmee et al. (2010) reported that multiple regression analysis is used as tools of
assessment of the effects of two or more independent factors on the dependent variables.
From the regression analysis carried out, the best performing functional model was developed. The
criteria for adjudging this model was the value of the coefficient of determination and adjusted
coefficient of determination. Models were checked for adequacy using these statistics and the one
found to be adequate was selected from among the other possible combinations of the models. The
coefficients of determination R² is a measure of the total variation of the observed values of the
extracted oil about the mean explained by the fitted model (Shridhar et al, 2010).
3.0 Results and Discussion
3.1 Moisture Content of the Neem Seeds
The mean initial moisture content of the neem seeds based on triplicate determinations was 8.1%.
This was further reconditioned to 6.3, 13.2 and 16.6%.
3.2 Initial Oil Content of Neem Seeds prior to Treatment
The average initial oil content of the seeds was 37.71%
3.3 Percentage Oil Yield of Treated Neem Seeds

Oil yield of the different treatment combinations was calculated and shown in Table 3.
Table 3: Results of Oil Yield
S/No Treatment combination Oil Yield
Replicate 1 Replicate 2 Replicate 3
1 m1t1p1β1 13.47 12.98 13.42
2 m1t1p1β2 13.35 13.37 13.34
3 m1t1p1β3 13.56 13.42 12.92
4 m1t1p1β4 13.72 13.55 12.96
5 m1t1p2β1 16.07 15.92 15.95
6 m1t1p2β2 15.98 15.94 16.03
7 m1t1p2β3 16.17 16.12 15.97
8 m1t1p2β4 16.21 16.02 16.29
9 m1t1p3β1 20.21 20.33 20.17
10 m1t1p3β2 20.72 20.67 20.83
11 m1t1p3β3 21.77 20.89 21.71
12 m1t1p3β4 22.55 21.97 22.32
13 m1t1p4β1 18.21 18.17 18.09
14 m1t1p4β2 19.01 18.67 18.98
15 m1t1p4β3 19.16 18.93 19.12
16 m1t1p4β4 20.47 19.87 20.22
17 m1t2p1β1 15.72 15.86 16.01
18 m1t2p1β2 17.12 16.97 17.09
455
19 m1t2p1β3 17.17 17.19 16.99

20 m1t2p1β4 17.22 17.37 17.09
21 m1t2p2β1 19.99 19.11 19.22
22 m1t2p2β2 20.02 20.61 20.99
23 m1t2p2β3 22.17 21.78 21.54
24 m1t2p2β4 21.82 22.35 22.77
25 m1t2p3β1 25.21 24.72 24.01
26 m1t2p3β2 26.77 25.91 26.74
27 m1t2p3β3 27.06 28.92 27.22
28 m1t2p3β4 29.74 29.61 29.42
29 m1t2p4β1 22.78 21.87 22.92
30 m1t2p4β2 23.11 23.09 23.21
31 m1t2p4β3 24.27 24.17 24.28
32 m1t2p4β4 24.98 23.44 24.67
33 m1t3p1β1 18.32 18.17 18.38
34 m1t3p1β2 20.11 19.96 20.28
35 m1t3p1β3 22.02 21.89 22.11
36 m1t3p1β4 23.13 23.06 23.15
37 m1t3p2β1 25.07 25.12 24.93
38 m1t3p2β2 26.82 27.04 26.63
39 m1t3p2β3 27.97 27.92 28.04
40 m1t3p2β4 28.53 28.64 28.43
41 m1t3p3β1 30.07 29.97 30.09
42 m1t3p3β2 31.81 31.72 31.95
43 m1t3p3β3 32.99 32.78 32.86
44 m1t3p3β4 34.17 33.92 34.09
45 m1t3p4β1 28.87 28.65 28.93
46 m1t3p4β2 29.35 29.41 29.22
47 m1t3p4β3 31.22 31.28 30.96
48 m1t3p4β4 31.72 31.59 31.52
49 m1t4p1β1 18.12 18.23 17.97
50 m1t4p1β2 18.17 18.09 18.24
51 m1t4p1β3 17.37 16.82 17.17
52 m1t4p1β4 17.99 18.04 17.87
53 m1t4p2β1 21.08 21.11 21.99
54 m1t4p2β2 21.97 21.72 21.54
55 m1t4p2β3 21.14 21.52 20.92
56 m1t4p2β4 21.32 21.41 21.96
57 m1t4p3β1 25.32 25.11 25.54
58 m1t4p3β2 24.87 25.22 24.98
59 m1t4p3β3 26.11 25.84 26.03
456
60 m1t4p3β4 26.20 26.12 26.29

61 m1t4p4β1 24.21 24.04 24.33
62 m1t4p4β2 23.98 23.27 23.72
63 m1t4p4β3 24.02 24.94 24.14
64 m1t4p4β4 25.22 26.13 25.18
65 m2t1p1β1 15.34 15.12 14.29
66 m2t1p1β2 16.21 16.01 15.72
67 m2t1p1β3 15.77 15.56 15.32
68 m2t1p1β4 16.01 16.54 15.82
69 m2t1p2β1 18.35 19.03 18.66
70 m2t1p2β2 20.98 21.04 20.75
71 m2t1p2β3 22.77 22.13 21.94
72 m2t1p2β4 23.96 23.07 24.15
73 m2t1p3β1 22.11 23.03 23.46
74 m2t1p3β2 23.72 23.92 23.73
76 m2t1p3β4 27.27 26.83 26.51
77 m2t1p4β1 20.16 21.72 20.91
78 m2t1p4β2 22.32 21.92 22.41
79 m2t1p4β3 24.97 23.18 23.24
80 m2t1p4β4 25.02 25.78 26.11
81 m2t2p1β1 18.33 17.62 17.97
82 m2t2p1β2 20.11 19.97 19.82
83 m2t2p1β3 20.79 21.83 22.92
84 m2t2p1β4 21.55 23.10 22.99
85 m2t2p2β1 23.65 23.33 23.72
86 m2t2p2β2 25.86 25.22 25.19
87 m2t2p2β3 27.09 26.84 26.53
88 m2t2p2β4 28.12 27.55 28.05
89 m2t2p3β1 28.88 29.21 28.83
90 m2t2p3β2 29.91 29.07 30.76
91 m2t2p3β3 30.74 29.83 30.62
92 m2t2p3β4 31.41 32.03 31.74
93 m2t2p4β1 25.23 26.08 25.17
94 m2t2p4β2 26.45 26.59 27.21
95 m2t2p4β3 26.07 25.67 27.08
96 m2t2p4β4 25.99 26.34 26.46
97 m2t3p1β1 20.27 19.86 20.52
98 m2t3p1β2 22.08 22.73 22.95
99 m2t3p1β3 22.98 23.51 23.83
100 m2t3p1β4 23.19 23.94 23.57
101 m2t3p2β1 26.63 25.83 26.22
457
102 m2t3p2β2 28.92 28.76 28.51

103 m2t3p2β3 30.92 31.08 29.43
104 m2t3p2β4 31.18 31.93 31.67
105 m2t3p3β1 32.77 33.12 32.19
106 m2t3p3β2 32.99 34.02 33.73
107 m2t3p3β3 34.28 35.11 34.02
108 m2t3p3β4 36.12 36.72 36.23
109 m2t3p4β1 28.11 28.93 28.72
110 m2t3p4β2 30.82 29.47 30.68
111 m2t3p4β3 32.98 31.33 31.54
112 m2t3p4β4 33.44 32.97 33.08
113 m2t4p1β1 19.22 19.17 19.11
114 m2t4p1β2 18.43 18.22 18.17
115 m2t4p1β3 17.88 18.13 17.54
116 m2t4p1β4 18.52 18.03 18.21
117 m2t4p2β1 21.17 22.02 21.34
118 m2t4p2β2 22.43 21.92 22.21
119 m2t4p2β3 22.98 21.63 22.42
120 m2t4p2β4 22.22 21.17 23.04
121 m2t4p3β1 26.82 26.11 26.49
122 m2t4p3β2 26.17 25.23 25.77
123 m2t4p3β3 27.01 26.15 27.22
124 m2t4p3β4 27.21 27.53 27.33
125 m2t4p4β1 25.26 25.72 24.98
126 m2t4p4β2 25.73 25.46 25.61
127 m2t4p4β3 26.12 25.34 26.07
128 m2t4p4β4 26.26 26.72 25.41
129 m3t1p1β1 12.71 12.32 12.46
130 m3t1p1β2 12.98 13.14 12.73
131 m3t1p1β3 13.33 13.49 13.22
132 m3t1p1β4 13.16 13.32 13.58
133 m3t1p2β1 16.23 16.17 16.62
134 m3t1p2β2 16.79 17.14 17.24
135 m3t1p2β3 17.92 17.63 17.51
136 m3t1p2β4 17.77 18.21 18.05
137 m3t1p3β1 19.51 19.24 19.46
138 m3t1p3β2 19.22 19.67 19.81
139 m3t1p3β3 21.45 20.04 20.52
140 m3t1p3β4 21.97 20.55 20.78
141 m3t1p4β1 17.63 17.82 17.07
142 m3t1p4β2 18.41 18.22 18.15
458
143 m3t1p4β3 19.02 18.51 18.73

144 m3t1p4β4 19.89 19.18 19.76
145 m3t2p1β1 15.63 15.32 15.41
146 m3t2p1β2 16.13 16.37 16.74
147 m3t2p1β3 17.27 16.85 17.13
148 m3t2p1β4 17.32 17.01 17.24
149 m3t2p2β1 19.16 19.21 19.07
150 m3t2p2β2 21.22 20.54 20.83
151 m3t2p2β3 22.17 21.26 21.12
152 m3t2p2β4 23.25 22.51 22.18
153 m3t2p3β1 24.84 24.11 25.17
154 m3t2p3β2 26.92 25.84 25.44
155 m3t2p3β3 28.55 27.92 27.19
156 m3t2p3β4 29.11 28.97 29.03
157 m3t2p4β1 21.43 20.87 21.91
158 m3t2p4β2 22.72 21.43 22.31
159 m3t2p4β3 23.97 23.19 23.46
160 m3t2p4β4 24.99 23.82 23.64
161 m3t3p1β1 17.54 17.88 17.32
162 m3t3p1β2 18.22 18.84 18.43
163 m3t3p1β3 19.76 19.36 20.08
164 m3t3p1β4 21.11 20.87 21.16
165 m3t3p2β1 23.71 23.44 22.95
166 m3t3p2β2 25.23 25.09 25.72
167 m3t3p2β3 26.19 26.76 26.50
168 m3t3p2β4 27.91 27.84 28.67
169 m3t3p3β1 30.04 30.25 30.61
170 m3t3p3β2 31.87 31.23 31.73
171 m3t3p3β3 32.18 32.64 32.81
172 m3t3p3β4 34.27 33.92 34.19
173 m3t3p4β1 22.74 21.71 22.15
174 m3t3p4β2 23.12 22.17 22.34
175 m3t3p4β3 24.21 24.04 23.61
176 m3t3p4β4 26.91 25.11 25.03
177 m3t4p1β1 16.74 16.23 16.89
178 m3t4p1β2 17.49 16.54 17.91
179 m3t4p1β3 17.23 18.02 17.68
180 m3t4p1β4 17.72 17.63 17.87
181 m3t4p2β1 19.75 19.22 19.03
182 m3t4p2β2 20.78 19.93 20.14
183 m3t4p2β3 21.09 20.73 20.91
459
184 m3t4p2β4 21.85 21.23 20.98

185 m3t4p3β1 23.23 22.72 23.46
186 m3t4p3β2 21.26 21.07 22.19
187 m3t4p3β3 21.41 22.72 22.29
188 m3t4p3β4 22.22 22.58 23.04
189 m3t4p4β1 23.10 23.21 23.53
190 m3t4p4β2 22.21 21.73 21.92
191 m3t4p4β3 22.02 21.61 21.43
192 m3t4p4β4 21.87 21.35 21.26
193 m4t1p1β1 9.85 10.12 9.64
194 m4t1p1β2 9.97 10.23 9.71
195 m4t1p1β3 9.72 10.76 10.53
196 m4t1p1β4 9.88 10.34 10.21
197 m4t1p2β1 11.24 11.71 12.07
198 m4t1p2β2 11.71 12.80 12.62
199 m4t1p2β3 11.98 12.93 12.44
200 m4t1p2β4 12.54 13.08 13.22
201 m4t1p3β1 14.27 15.16 14.83
202 m4t1p3β2 14.75 15.22 15.60
203 m4t1p3β3 14.99 15.77 17.03
204 m4t1p3β4 16.22 16.31 17.07
205 m4t1p4β1 13.01 13.87 13.46
206 m4t1p4β2 13.33 13.65 14.12
207 m4t1p4β3 14.12 14.08 14.81
208 m4t1p4β4 14.87 15.24 15.19
209 m4t2p1β1 10.11 10.94 10.72
210 m4t2p1β2 11.27 12.01 11.97
211 m4t2p1β3 11.82 11.91 12.74
212 m4t2p1β4 12.01 12.18 12.50
213 m4t2p2β1 13.67 13.22 13.81
214 m4t2p2β2 15.32 15.19 14.28
215 m4t2p2β3 17.12 16.91 16.56
216 m4t2p2β4 17.94 17.22 17.67
217 m4t2p3β1 16.17 16.67 16.01
218 m4t2p3β2 18.79 17.34 18.73
219 m4t2p3β3 20.22 19.11 20.76
220 m4t2p3β4 21.17 21.83 22.71
221 m4t2p4β1 15.74 15.22 15.51
222 m4t2p4β2 17.61 16.91 16.34
223 m4t2p4β3 19.37 18.17 18.76
Where: m=moisture content, t=temperature, p=pressure, t= heating duration
460
3.4. Analysis of Variance of oil yield

The results for the analysis of variance of oil yield is presented in Table 4.
Table 4: Analysis of Variance for Oil Yield
Partial Eta
Source Sum of Squares Df Mean Square F Sig. Squared
M 8785.57 3.00 2928.53 15700.00 0.000 0.989
ϴ 6484.75 3.00 2161.58 11590.00 0.000 0.985
P 7935.16 3.00 2645.05 14180.00 0.000 0.988
638.89 3.00 212.96 1141.00 0.000 0.870
mϴ 327.47 9.00 36.39 195.02 0.000 0.774
Mp 172.45 9.00 19.16 102.70 0.000 0.644
ϴp 355.41 9.00 39.49 211.65 0.000 0.788
ϴt 216.41 9.00 24.05 128.88 0.000 0.694
mϴp 273.38 27.00 10.13 54.27 0.000 0.741
Error 95.53 512.00 0.19
Total 25285.01 587.00
a. R Squared = .996 (Adjusted R Squared = .994)
b. Ftab= 3.84 5% level of Significance
*m=moisture content, ϴ=temperature, p=pressure, t= heating duration
From Table 4, it can be inferred that the process conditions (moisture content, heating temperature,
pressure and heating duration) and their interactions had significant effect on oil yield at 5% level
of significance. This implies that at least one of the mean treatment effects is significantly
different from the others. The Partial Eta Square (PES) statistics shows the ‘practical’significance
of each process condition. Larger values of PES indicate a greater amount of variation accounted
for by the model process conditions to a maximum of 1, i.e. the closer value of PES of process
condition is to 1, the higher the contribution or effect of such term to the significance of the model.
Hence it can be concluded from Table 4 that all the process conditions and their interactions
contributed strongly to the significance of the model.
3.5 Effect of moisture content on oil yield
Duncan’s New Multiple Range Test (DNMRT) was conducted to determine the differences in the
mean treatment effect of moisture content on oil yield as reflected in the ANOVA table (Table 4)
and this is shown in Table 5.
Table 5: Duncan’s New Multiple Range Test for Oil Yield at various Moisture Contents
Moisture content (% wb) Oil yield (%)
16.6 15.62a
13.2 21.21b
6.3 22.37c
8.1 24.86d
Means with the same alphabet are not significantly different from each other
461
Table 5 showed that seeds with higher moisture content yielded less oil as compared to seeds with
lower moisture content. This agrees with the findings by Farsie and Singh (1985) who reported
maximum oil recovery for sunflower seeds expressed at 6% moisture content. The authors reported
further that increasing the moisture content to 14% decreased the oil recovery by 16%. However,
from Table 5, it is observed that moisture contents at each level recorded significantly different oil
yields. The maximum oil yield of 24.86% was observed at 8.1% mc while the minimum oil yield
of 15.62% was recorded at the highest moisture content of 16.6%. Considering only the single
effect of moisture content on oil yield. Both two-way and three way interactions of the process
conditions were also significant indicating that the process conditions are highly interactive in
nature.
3.6 Effect of heating temperature on oil yield
It can be inferred from Table 4 (ANOVA on oil yield) that heating temperature and its interactions
with moisture content, pressure and heating duration had significant effect on oil yield at 5% level
of significance. It implies that they all contributed significantly to the oil yield obtained.
The result of the Duncan’s New Multiple Range Test (DNMRT) which was conducted to
determine the differences in the mean treatment effect of heating temperature on oil yield is shown
in Table 6.
Table 6: Duncan’s New Multiple Range Test for Oil Yield at various Temperatures
Temperature Oil yield
55 17.24a
100 20.06b
70 21.44c
85 25.31d
Means with the same alphabet are not significantly different from each other
Table 6 (oil yield at various temperatures) which shows the comparison among the four levels of
heating temperature revealed that at any particular heating temperature, the observed means of oil
yield are significantly different from each other. Table 6 further revealed that high oil yield was
obtained at 85°C while the lower oil yield was obtained at 55°C. Considering only the single effect
of temperature. This trend compares favourably with previous researches on oil expression by
Adeeko and Ajibola (1990), Fasina and Ajibola (1989) Ajibola et al. (1990b), Fasina and Ajibola
(1990), and Tunde-Akintundeet al. (2001). According to Adeeko and Ajibola (1990); Fasina and
Ajibola (1989), heat treatment of oilseeds prior to oil extraction ruptures oil cells, reduces oil
viscosity and increases flowability. These processes increase oil yield and extraction efficiency by
allowing oil to flow easily during extraction (Olaniyan, 2010b); thus, this explains the increase in
oil yield as the heating temperature was increased from 55 to 85°C in this study. Ayenew (2000)
reported that oil recovery from Niger seed samples were influenced by increasing heat levels, and
that the optimum oil recovery was obtained when the Niger seed samples were heated at 900C
prior to expression.
Soetaredjo et al. (2008) however reported that preheating of neem seeds prior to oil expression
resulted in lower yield of oil. They showed that as the preheating temperature increased, the yield
of neem oil decreased, although they reported that this decrease was mainly related to the moisture
462
content and structure of the seeds. Adejumo et al. (2013) reported the highest oil yield of 33.7%
for moringa oleifera seed oil when the extraction was done at 100°C as against the 32.2 and
30.90% obtained at 130 and 150°C respectively.
3.7. Effect of heating duration on oil yield
From the result of the Analysis of variance for oil yield presented in Table 4, it can be inferred that
heating duration and its interaction with heating temperature had significant effect on oil yield at
5% level of significance. The result of the Duncan’s New Multiple Range Test (DNMRT) which
was conducted to determine the differences in the mean treatment effect of heating duration on oil
yield is shown in Table 7.
Table 7: Duncan’s New Multiple Range Test for Oil Yield at various Heating Durations
Heating duration Oil yield
5 19.77a
10 20.64b
15 21.42c
20 22.23d
Means with different alphabets are significantly different from each other
Table 7 (oil yield at various heating durations) which shows the comparison among the four levels
of heating duration revealed that mean oil yields at the different levels were significantly different
from each other. Longer heating duration yielded more oil as compared to shorter heating duration.
This suggests that for more oil yield, the heating duration should be increased as much as possible,
provided that other important qualities of the oil are not compromised. The highest heating
duration (20 min) for this experiment recorded the highest oil yield of 22.22% which is statistically
higher than all other heating durations used in this experiment. Considering only the single effect
of heating duration
Olaniyan (2011) reported that maximum oil yield was extracted from orange seeds at 20 min
roasting duration which was the highest duration used in the extraction process. Tunde-Akintunde
et al. (2001) reported that oil yield of mechanically expressed soybean oil increased with increase
in heating time from 15 to 30 min; the authors observed that the highest oil yield of 10.4% was
obtained when the samples were heated for 30 min at a temperature of 80°C. Manpouya et al.
(2013) and Jing et al. (2012) observed that extraction time of 2 and 3 h gave optimum oil yield in
the solvent extraction of safou pulp (Dacryodes Deulis) and rape seed oils. Awolu et al. (2013)
reported that increasing the extraction time and particle size of neem seed mass from 1 h and 0.42
mm to 2 h and 1.39 mm, respectively, gave maximum oil yield in the solvent extraction of oil from
neem seed.
Adeeko and Ajibola (1990) reported that the rate of oil expression from groundnut seeds increased
with increase in heating duration and heating temperature when moisture content and pressing pressure
are held constant. Bamgboye and Adejumo (2011) reported an increase in rosselle oil yield as the
heating time was increased from 15 min to 25 min.
3.8 Effect of machine pressure on oil yield

From the ANOVA table for oil yield presented in Table 4, it can be inferred that applied pressure
and its interactions had significant effect on oil yield at 5% level of significance. The partial eta
463
squared (PES) statistics reported the ‘practical’ significance of this process condition. Hence, it
was concluded based on data from Table 4 that applied pressure and its interactions contributed
strongly to the significance of the process. Duncan’s New Multiple Range Test (DNMRT) was
therefore conducted to determine the differences in the mean treatment effects of applied pressure
on oil yield (Table 8).
Table 8: Duncan’s New Multiple Range Test for Oil Yield at various Pressures
Pressure (KN/m2) Oil yield (%)
602.91 16.18a
723.07 20.52b
1334.88 22.38c
925.84 24.99d
Means with different alphabets are significantly different from each other
The result of the comparison among the four levels of applied pressure revealed that each level of
pressure recorded significantly different oil yield. Higher pressure seems to yield higher oil and
becoming a maximum at 925.84 KN/m2. Further increase in applied pressure beyond this point
(925.84 KN/m2) led to a reduction in yield. The highest oil yield of 24.99% was recorded at 925.84
KN/m2 while the minimum oil yield of 16.18% was observed at 602.91 KN/m2.
Soetaredjo et al. (2008) reported that seed particles will deform and compactly fill up empty voids
at low pressure level, but when the pressure is increased, the voids will diminish and the seed
particles begin to resist the applied pressure through contact points between particles. Further
increase in pressure will force the oil to start flowing out of the particles and the authors concluded
that the neem oil point appeared to be at 13789.51 KN/m2, and that 34473.79 KN/m2 is the
optimum pressure since further pressure beyond this ( such as 41638.54 KN/m2) gave insignificant
increase on the oil yield. Adeeko and Ajibola (1990) reported that this was so because empty voids
between particles from which the oil could flow out were becoming smaller at higher pressures.
Olaniyan and Oje (2007) reported that oil yield of shea butter from shea kernel increased
progressively with increase in applied pressure from 1.5 to 8.8 MPa. Ebewele et al. (2010) also
reported a consistent increase in oil yield of rubber seed oil when the pressure was increased from
5 MPa to 8 MPa.
The trend observed in this study also agrees with the work of the following researchers: Adeeko
and Ajibola (1990), who reported that Oil yield from groundnut seeds increased with increase in
pressure up to 20 MPa beyond which the yield levelled off; Ajibola et al. (1990a) reported a
significant increase in oil yield from melon seeds when applied pressure was increased from 5 to
18 MPa but oil yield either levelled off or decreased slightly when the pressure was increased to 25
MPa; Fasina and Ajibola (1989) reported an increase in oil yield from conophor nuts as pressure
increased from 10 MPa to 25 MPa. This behaviour can be explained as follows: During the process
of oil expression from oilseeds, increasing the pressure applied during screw pressing tends to
decrease the size of the capillaries through which oil flows and further increase in pressure may
eventually lead to the sealing of the capillaries and some inter kernel voids (Ward, 1976; Adeeko
and Ajibola, 1990).
464
3.9 Model Developed

The coefficients of determination, R² for the response, yield, was 0.907. The model relating the
process condition and output is given below:
Oil Yield
ϴ
*ϴ= Temperature, m= moisture content, t= heating duration, p= pressure

From the model equation generated, the following can be deduced:
Temperature (ϴ)
Temperature had positive linear effects on oil yield. It also had positive quadratic effects on oil
yield. The interactions between temperature and moisture content, temperature and pressure,
temperature and heating duration had positive effect on oil yield.
Heating Duration (t)
Heating duration had positive linear effect on oil yield
Moisture Content (m)

Moisture content had positive linear effect on oil yield; it had negative quadratic effect on oil yield.
The interactions between moisture content and pressure and moisture content and temperature had
negative effects on oil yield.
Pressure (p)
Pressure had positive linear effect on oil yield; it had negative quadratic effect on oil yield.
Regression statistics was used to check the model adequacy. The adjusted coefficient of
determination which defined the percentage of total variability explained by the models was
90.4%, oil yield. This high percentage of total variability explained by the model implies good fit.
The values of coefficient of determination (R2, 90.7%) and adjusted coefficient of determination
(90.4%) are relatively close for all process parameters. According to David et al. (1998), this is
what should be expected of good models.
4. Conclusions
The data generated from this study have been used to show that temperature, heating duration,
moisture content, pressure and their interactions all proved to determine the quantity of oil yield
from neem seed kernels. The mathematical model developed is adequate to relate the process input
to the output. The reliability and adequacy of the model developed to do this was examined using
the coefficient of determination R2 and adjusted coefficient of determination (R2). The results
obtained indicated that the model was statistically adequate and can be used to relate process input
to process output. The coefficient of determination and adjusted coefficient of determination (R2)
values were quite high (90.7 and 90.4%), this high percentage of total variability indicate goodness
of fit.
References
Abdullahi, YN. 2004. Medicinal Uses of Neem. National Research Institute on Chemical
Technology (NARICT) Seminar/ Exposition: Neem for Economic Empowerment. pp 1-7.
465
Adeeko, KA. and Ajibola, OO. 1990. Processing Factors Affecting Yield and Quality of
Mechanically Expressed Groundnut Oil. Journal of Agricultural Engineering Research;
45(1):31-43.
Adejumo, BA., Alakowe, AT and Obi, DE. 2013. Effect of Heat Treatment on the Characteristics
and Oil Yield of MoringaOleifera Seeds. The International Journal of Engineering and Science
(IJES); 2(1): 232-239.
Adewoye, TL. and Ogunleye, OO. 2012. Optimization of Neem Seed Oil Extraction Process
Using Response Surface Methodology. Journal of Natural Sciences Research; 2 (6):66-76.
Ajibola, OO., Eniyemo, SE., Fasina, OO. and Adeeko, KA. 1990a. Mechanical Expression of Oil
from Melon Seeds. Journal of Agricultural Engineering Research; 45:45 –53.
Ajibola, OO., Bakare, FA. and Fasina, OO. 1990b. Effects of some Processing Factors on Yield of
Oil Expressed from Rubber Seeds. Ife Journal of Technology; 2(2): 1-7.
Akinoso, R. 2006. Effects of Moisture Content, Roasting Duration and Temperature on Oil Yield
and Quality of Palm Kernel (Elaeisguineensis) and Sesame (Sesamium indicum) Oils. Ph.D Thesis,
Department of Agricultural and Environmental Engineering, University of Ibadan, Ibadan, Nigeria.
Akinoso, R., Igbeka, JC., Olayanju, T. and Bankole, L. 2006a. Modeling of Oil Expression from
Palm Kernel (ElaeisguineensisJacq.). Agricultural Engineering International:CIGR Ejournal; 8: 1-
7.
ASABE. 2008. Method of Determining and Expressing the Fineness of Feed Materials by Sieving,
ASABE S319. American Society of Agricultural and Biological Engineers, 2950 Niles Road, St.
Joseph MI 49085-9659, Michigan, United States of America.
Awolu, OO., Obafaye, RO. and Ayodele, BS. 2013. Optimization of Solvent Extraction of Oil
from Neem (Azadirachta indica) and its Characterizations. Journal of Scientific Research and
Reports; 2(1).Article no: JSRR.2013.020.
Ayenew, M. 2000. Processing of Niger Seed in Small Mechanical Expellers as Affected by Post
Harvest Storage and Pre-Extraction Treatments. Agricultural Mechanization in Asia, Africa and
Latin America; 31(4): 62-66.
Bamgboye, IA. and Adejumo, OI. 2011. Effects of Processing Parameters of Roselle Seed on It’s
Oil Yield. International Journal of Agricultural and Biological Engineering, IJABE; 4(1):83-86.
Boonmee, K., Chuntranuluck, S., Punsuvon, V. and Silayoi, P. 2010. Optimization of Biodiesel
Production from Jatropha (Jatropha Curcas L.), using Surface Response Methodology. Kasetsart
Jounal of Natural Sciences; 44: 290-299.
David, D., Steppan, JW. and Robert, PY. 1998. Essential Regression and Experimental Design for
Chemists and Engineers, a Manual for Regression Models. pp. 12-88.
466
Ebewele, RO., Iyayi, AF. and Hymore, FK. 2010. Considerations of the Extraction Process and
Potential Technical Applications of Nigerian Rubber Seed Oil. International Journal of the
Physical Sciences; 5(6): 826-831.
Farsie, A. and Singh, MS. 1985. Energy Models for Sunflower Oil Expression. Transactions of the
American Society of Agricultural Engineers; 28: 275 –279.
Fasina, OO. and Ajibola, OO. 1989. Mechanical Expression of Oil from Conophor Nut
(Tetracarpidium conophorum). Journal of Agricultural Engineering Research; 46(45-53).
Fasina, OO. and Ajibola, OO. 1990. Development of Equations for the Yield of Oil expressed
from Conophor Nut. Journal of Agricultural Engineering Research; 46: 45-53.
Ikasari, M. and Indraswati, NA. 2008. Physical Properties of Foods and Processing Systems.
Chichester: Ellis Harwood.
Jing, Y., Jing, W., Chunming, L., Zhiqiang, L. and Qiang, W. 2012. Application of Response
Surface Methodology to Optimal Supercritical Carbondioxide Extraction of Oil from
Rapeseed (Brassica napum L.). International Journal of Food Science and Technology;
47(6):1115-1121.
Khan, LM. and Hanna, MA. 1983. Expression of Oil from Oilseeds –A Review. Journal of
Agricultural Engineering Research; 28:495-503.
Mampouya, D., Kama, NR., Goteni, S., Loumouamou, AN., Kinkela, T. and Silou, T. 2013.
Optimization of the Soxhlet Extraction of Oil from Safou Pulp (Dacryodes Deulis). Advance
Journal of Food Science and Technology; 5(3):230-235.
Olajide, JO. 2000. Process Optimisation and Modelling of Oil Expression from Groundnut and
Sheanut Kernel. Ph.D Thesis, University of Ibadan, Nigeria.
Olaniyan, AM. and Oje, K. 2007. Development of Mechanical Expression Rig for Dry Extraction
of Shea Butter from Shea kernel. Journal of Food Science and Technology; 44(5): 465-470.
Olaniyan, AM. 2010b. Effect of Extraction Conditions on the Yield and Quality of Oil from Castor
Bean. Journal of Cereals and Oilseeds; 1(2):24-33.
Olaniyan, AM. and Oje, K. 2011. Development of Model Equations for Selecting Optimum
Parameters for Dry Process of Shea Butter Extraction. Journal of Cereals and Oilseeds; 2(4): 47-
56.
Oyinlola, A. and Adekoya LO. 2004. Development of a Laboratory Model Screw Press for Peanut
Oil Expression. Journal of Food Engineering; 64:221-227.
Rajeev S. 2009. Dinkal Agro Inc: Organic for Healthy Living, Dehra Dun, India: International
Book Distributors. pp123 – 128.
Sablani, SS, Datta, AK, Rahman, MS. and Mujumdar, AS. 2006. Handbook of Food and
Bioprocess Modeling Techniques. CRC Press. Pp 624.
467
Shridhar, BS., Beena, KV., Anita, MV. and Paramjeet, KB. 2010. Optimization and
Charaterization of Castor Seed Oil. Leonardo Journal of Science; 17:59-70.
Soetaredjo, FE., Budijanto, GM.; Prasetyo, RI. and Indraswati, N. 2008. Effects of Pre- Treatment
Condition on the Yield and Quality of Neem Oil Obtained by Mechanical Pressing. ARPN Journal
of Engineering and Applied Sciences; 3(5):45-49.
Torres, MM. and Maestri, DM. 2006. Chemical Composition of Arbequina Virgin Olive Oil in
Relation to Extraction and Storage Conditions. Journal of the Science of Food and Agriculture;
86(14): 2311–2317.
Tunde-Akintunde, TY. 2000. Predictive Models for Evaluating the Effectiveness of some
Parameters on Yield and Quality of some Soybean Products. Ph.D thesis, University of Ibadan,
Nigeria.
Tunde-Akintunde, TY., Akintunde, BO. and Igbeka, JC. 2001. Effect of Processing Factors on
Yield and Quality of Mechanically Expressed Soybeans Oil. Journal of Agricultural Engineering
Technology; 55, 86-92.
Ward, JA. 1976. Processing High Oil Content Seeds in Continuous Screw Press. Journal of
American Oil Chemist Society; 53: 261 –264.
468

Statistical Modelling of Oil Expression From Neem Seed Using A Screw Press

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Statistical Modelling of Oil Expression From Neem Seed Using A Screw Press

Uploaded by

Copyright:

Available Formats

Arid Zone Journal of Engineering, Technology and Environment, September, 2018; Vol.

STATISTICAL MODELLING OF OIL EXPRESSION FROM NEEM SEED USING A

Keywords: Neem seed, Oil, Factorial design, Moisture content, Temperature

Fig 1: NCRI, Badeggi Developed Oil Expeller

2.2.2 Moisture Content Determination

2.3 Statistical Analysis

3.3 Percentage Oil Yield of Treated Neem Seeds

19 m1t2p1β3 17.17 17.19 16.99

60 m1t4p3β4 26.20 26.12 26.29

102 m2t3p2β2 28.92 28.76 28.51

143 m3t1p4β3 19.02 18.51 18.73

184 m3t4p2β4 21.85 21.23 20.98

3.4. Analysis of Variance of oil yield

3.8 Effect of machine pressure on oil yield

3.9 Model Developed

*ϴ= Temperature, m= moisture content, t= heating duration, p= pressure

Moisture Content (m)

You might also like